Phylodynamics unveils invading and diffusing patterns of dengue virus serotype-1 in Guangdong, China from 1990 to 2019 under a global genotyping framework

Abstract

Background: The strong invasiveness and rapid expansion of dengue virus (DENV) pose a great challenge to global public health. However, dengue epidemic patterns and mechanisms at a genetic scale, particularly in term of cross-border transmissions, remain poorly understood. Importation is considered as the primary driver of dengue outbreaks in China, and since 1990 a frequent occurrence of large outbreaks has been triggered by the imported cases and subsequently spread to the western and northern parts of China. Therefore, this study aims to systematically reveal the invasion and diffusion patterns of DENV-1 in Guangdong, China from 1990 to 2019.

Methods: These analyses were performed on 179 newly assembled genomes from indigenous dengue cases in Guangdong, China and 5152 E gene complete sequences recorded in Chinese mainland. The genetic population structure and epidemic patterns of DENV-1 circulating in Chinese mainland were characterized by phylogenetics, phylogeography, phylodynamics based on DENV-1 E-gene-based globally unified genotyping framework.

Results: Multiple serotypes of DENV were co-circulating in Chinese mainland, particularly in Guangdong and Yunnan provinces. A total of 189 transmission clusters in 38 clades belonging to 22 subgenotypes of genotype I, IV and V of DENV-1 were identified, with 7 Clades of Concern (COCs) responsible for the large outbreaks since 1990. The epidemic periodicity was inferred from the data to be approximately 3 years. Dengue transmission events mainly occurred from Great Mekong Subregion-China (GMS-China), Southeast Asia (SEA), South Asia Subcontinent (SASC), and Oceania (OCE) to coastal and land border cities respectively in southeastern and southwestern China. Specially, Guangzhou was found to be the most dominant receipting hub, where DENV-1 diffused to other cities within the province and even other parts of the country. Genome phylogeny combined with epidemiological investigation demonstrated a clear local consecutive transmission process of a 5C1 transmission cluster (5C1-CN4) of DENV-1 in Guangzhou from 2013 to 2015, while the two provinces of Guangdong and Yunnan played key roles in ongoing transition of dengue epidemic patterns. In contextualizing within Invasion Biology theories, we have proposed a derived three-stage model encompassing the stages of invasion, colonization, and dissemination, which is supposed to enhance our understanding of dengue spreading patterns.

Conclusions: This study demonstrates the invasion and diffusion process of DENV-1 in Chinese mainland within a global genotyping framework, characterizing the genetic diversities of viral populations, multiple sources of importation, and periodic dynamics of the epidemic. These findings highlight the potential ongoing transition trends from epidemic to endemic status offering a valuable insight into early warning, prevention and control of rapid spreading of dengue both in China and worldwide.

Keywords: China; Dengue serotype-1; Epidemic periodicity; Genetic population structure; Global genotyping framework; Molecular epidemiology; Phylodynamics; Phylogeography.

Fig. 1

The population structure of DENV-1 circulating in Chinese mainland was designated through the E-gene-based global genotyping framework of DENV-1 [5]. a Maximum likelihood phylogenetic analyses show that there are 38 clades labeled in color circulating in Chinese mainland, while the other clades are labeled in gray. b Histogram of DENV-1 clades circulating in Chinese mainland from 1990 to 2019. The inset is a partial enlargement view during 1990‒2012. c, d Distribution of reported DENV-1 strains in Chinese mainland and Guangdong Province, respectively. Pies are slices colored by the composition of DENV-1 clades

Fig. 2

DENV-1 epidemic periodicity in Chinese mainland represented by the dynamics of the number, growth rates, and estimated of multiannual periodicities of the cases, sequences, transmission clusters, and clades. a-d Yearly time series of the number of cases (a), sequences (b), transmission clusters (c), and clades (d). e–h Yearly time series of the growth rates of the number of cases (e), sequences (f), transmission clusters (g), and clades (h). g-i Estimates of multiannual periodicities of the number of cases (i), sequences (j), transmission clusters (k), and clades (l) using the models driven by simple seasonal sine curves. Edge effects in the wavelet transforms may influence results before and after the purple-gray shadow in i, j, k, and l

Fig. 3

Introducing and diffusing routes of the seven COCs including 1E1, 1H4, 1J7, 1K1, 1L1, 1L2, and 5C1 in Chinese mainland inferred by BSSVS. a Phylogeographic reconstruction of the introduction and diffusion events of the seven COCs in Chinese mainland. Myanmar, Vietnam, Thailand, Singapore, Malaysia, Indonesia, and India, as well as Guangzhou, Zhongshan, Foshan, and Chaozhou in China are highlighted in red bold font as these countries/cities occurred more than ten migration events. b Dynamics and trends of the epidemic trunks of the seven COCs including 1E1, 1H4, 1J7, 1K1, 1L1, 1L2, and 5C1, representing the proportion of PACT-inferred dominant epidemic countries/cities changed dynamically along with different years

Fig. 4

Invasion and diffusion of COC 5C1 of DENV-1 in Chinese mainland. a Maximum clade credibility phylogeny of COC 5C1 based on the genomes. According to the topological structure and bootstrap values of the E-gene-based tree, COC 5C1 was further divided into 23 transmission clusters as showed in Chinese mainland showing in Additional file 10: Figure S6g. The present genome-based phylogenetic tree has the same clustering recognition ability as the E-gene-based tree, but it has significantly better resolution for discriminating the exact transmission process of each cluster. The critical SNPs, representing the evolution of clusters during their transmission process, are indicated by arrows at the cluster differentiation nodes on the left of the phylogenetic tree. Taking SNP 2359 as an example, it is T in the genomes of the upper branch, while it is C in the genomes of the lower branch. b Temporal dynamics of the designated transmission clusters of 5C1 based on its importation source and year. The inserted pie chart shows the composition ratio of 5C1 clusters with different size. c Genome-wide SNPs analysis of the lineage of 5C1 traced from Singapore. Only the positions with specific SNPs found in the strains of cluster 5C1-CN4 isolated in Chinese mainland were shown, with the SNPs of non-synonymous (NS) in blue and synonymous (S) in red. d Reconstructed genome-based transmission chain of 5C1-CN4 circulating in Chinese mainland. e. Spatiotemporal distribution of the indigenous cases observed in Tangjing Street, Guangzhou, China from Nov. 27th, 2014 to Apr. 4th, 2015. The cases reported in different month are shown in different color. The first indigenous dengue case (2015/KT827378) reported in Guangdong Province, China on Apr. 4th, 2015, is shown in red

Fig. 5

Invasion and diffusion patterns of DENV based on Invasion Biology theories demonstrated in this study. a Scheme of three-stages of invasion, colonization, and dissemination of DENV. The white areas indicates lands, and the blue areas indicates oceans. The circles represent the cities, where those in purple indicates the cities are with DENV endemic area or has been transferred to the endemic area, while those in green indicate the cities are without DENV endemic (dengue epidemics driven by the imported cases). The solid circle indicates the presence of large epidemic outbreaks, the hollow circle indicates the presence of sporadic cases, and the dashed circle indicates no reported DENV cases in the city. The purple direction line indicates the importation process from DENV endemic city to the non-endemic. The green direction line indicates the diffusion process of DENV between the non-endemic cities. The red arc arrow indicates that DENV survived the winter in the city and caused a dengue epidemic in the following year. As a result, these cities will be thrust into the ongoing high-risk transition process from epidemic to endemic through DENV’s continuing adaptation. b The schematic depicting the population dynamics of DENV during the importation and diffusion process, under the contextualization with the founder effects and population bottlenecks of mosquito-borne arboviruses, developed from Weaver et al. [32]